Compression and Noise Reduction of Biomedical Signals by Singular Value Decomposition
Thomas Schanze · IFAC-PapersOnLine · 2018
Compressing and denoising signals is important in signal processing. We introduce a method that bijectively maps a signal vector into a matrix. This matrix is decomposed by SVD into singular values and vectors to construct rank-one matrices. The application of the inverse mapping to the singular value related summation of these matrices yields a noise reduced signal vector. The scree test is used to select the rank-one matrices. Efficient storage of these signal driven matrices is realized by storing related singular values and vectors. The application of the method to simulated and biomedical signals shows its potential.